Facebook Questions Refreshed as Recommendation Engine
Facebook began rolling out Facebook Questions out to its U.S. users March 24, showing that the service has evolved into more of a recommendation engine to help friends answer friends' questions.
Facebook last July began testing Questions as an expert question-and-answer service. Users asked questions and waited for answers.
Users would click an "Ask Question" button at the top of the homepage, or ask questions about friends from their profiles right within the context of Facebook's Walls. When a question was posed, users could add a photo or poll.
Facebook patterned the service after Quora, an impersonal Q&A
startup created by former Facebook engineers Adam D'Angelo and Charlie
Facebook Questions Product Manager Adrian Graham said the social network quickly realized that on a social network, users' friends are often the best source of advice for where a user should go to dinner, what music they might like or even how to buy a car.
"We noticed that people were frequently asking for opinions ("what are your favorite restaurants in New York?") or hoping to learn about their friends ("what was your favorite movie as a kid, something you watched over and over?")," Graham wrote in a blog post.
"For most of these questions, experts weren't going to be the best source for advice. The answers to these questions are meaningful or interesting because you know your friends and your friends know you."
The new Facebook Questions, which is still public to all, helps users agree with an answer by checking a box, or adding a different response. Also, when a user answers a question, their friends can answer it, too, making it inherently more social. All answers are listed on the question's Web page.
Graham said for less personal questions seeking advice, users can still entertain advice from a broader group of people.
But to keep it crisply relevant, the Questions algorithm filters the answers to show users what their friends think first.
There are a lot of recommendation engines, from Hunch to Google's Aardvark, but none of these are paired with a social network of 600 million-plus users.
Amazon.com might have the largest scaling recommendation engine -- Netflix's isn't bad either -- but those services rely on existing user behaviors for recommendations. Facebook Questions is supercharged to be social; it remains to be seen whether people use it.